Deepfake Job Candidates: How to Spot Fake Video Interviews

This Isn't a Hypothetical

Remote hiring dropped the one thing that used to make impersonation hard: being in the room. Generative AI dropped the other thing — skill and cost. Put those together and you get organized networks of fake candidates running real-time face-swap video and cloned voices through legitimate interview pipelines, often targeting roles with access to financial systems, credentials, or sensitive data.

This shows up most in remote technical, financial, and systems-access roles — anywhere the payoff for a fake identity getting hired is worth the effort of building one.

What a Deepfake Interview Actually Looks Like

Two methods show up repeatedly:

Signs to Watch For

During the live interview:

• Blur or lag specifically around the face, while the background and hands stay sharp
• Lighting on the face that doesn't match the rest of the room
• Blinking that's too regular, too infrequent, or slightly out of sync with speech
• Audio that doesn't quite line up with lip movement on quick responses
• A candidate who won't turn their head or adjust the camera angle when asked
• Vague answers about their own resume — like they're reciting a script, not recalling experience

Across the hiring process:

• ID documents that don't clearly match the live video face on close comparison
• A resume with an unusually generic, hard-to-verify employment history
• Reluctance to appear on video for a final round after passing earlier screens
• References that are difficult to verify through official channels

What to Do When Something Feels Off

1. Ask for simple live actions a real-time deepfake struggles to render well: turn the head to the side, hold a hand briefly in front of the face, move closer to the camera
2. Record interviews with consent, so a suspicious round can be reviewed afterward instead of judged in the moment
3. Run the recording through an AI deepfake detection tool built for live meetings before final decisions on sensitive roles
4. Verify ID documents against live video, not just a static photo submitted separately
5. Cross-check employment history independently instead of relying only on candidate-provided references

Why a Detection Tool Beats a Gut Feeling

A well-executed deepfake is specifically designed to look normal to a person watching a video call — that's the entire point of the technology. An interviewer's instinct catches the sloppy fakes and misses the good ones. A dedicated video detection tool adds analysis a person can't do unaided: frame-by-frame artifact checks, audio-lip sync analysis, and a confidence score instead of a hunch.

This isn't a replacement for judgment — it's a filter. It tells you which interviews deserve a second look before you extend an offer or grant system access. For broader identity and fraud checks across the hiring pipeline, the same approach applies to detecting AI-driven scams and identity fraud more generally.